The goal of the somatic hypermutation group is to unveil the mechanism of immunoglobulin hypermutation that causes mutations to be introduced into the immunoglobulin locus during an immune response. Because somatic hypermutation contributes significantly to affinity maturation of B cells and thus to the development of high-affinity antibodies, we are also interested in the role of this mechanism in the development of highly autoreactive antibodies that contribute to the development of autoimmune diseases.[unreadable] [unreadable] During the course of an immune response, activated B cells either differentiate into antibody-forming cells (AFC) or seed germinal centers. AFC constitute the bulk of the primary immune B-cell response and are characterized by IgM antibodies early on, with some switching to downstream isotypes such as IgG late in a persistent infection. Germinal center B-cells, on the other hand, contribute little to the primary response: their main role is to undergo a process of hypermutation aimed at changing the specificity of the immunoglobulin receptor, and affinity maturation. Hypermutation is coupled to a cellular selection mechanism for high-affinity variants and this coupling leads to the generation of high-affinity memory B cells. Memory B cells contribute greatly to the secondary immune response and are, along with the memory T cell response, the reason why we are less likely to experience major symptoms of infection upon re-exposure to a particular antigen. The molecular basis of the immunoglobulin hypermutation mechanism is unknown and is one of the preoccupations of this laboratory. The cytosine deaminase, AID, which was recently found to play a pivotal role in immunoglobulin hypermutation, triggers this mechanism by deaminating cytosines in DNA, leading to the formation of uracil in the DNA. Uracil in DNA can be dealt with by uracil DNA glycosylases, or it is read as thymine during replication, leading to G:C to A:T transitions. Uniquely in immunoglobulin hypermutation, the evidence suggests that the presence of the G:U mismatch or the G:Ap site (Ap = abasic) leads to the recruitment of translesion synthesis DNA polymerases Eta and Zeta, leading to the generation of mutations at all bases, not just the G:C targeting expected from cytosine deamination on both strands. The repeated introduction of uracil into the DNA of immunoglobulin genes by AID and the resulting recruitment of the error-prone polymerases leads to a mutation frequency that is more than 1 million fold over background. Earlier, we found evidence that AID targets the immunoglobulin hotspot, DGYW, and demonstrated that AID is regulated by a nuclear export mechanism. However, we also demonstrated that AID does not intrinsically localize to the Ig locus but that it probably requires interaction with a protein partner for efficient localization and DNA binding. In another set of experiments, we have demonstrated that AID can bind and deaminate single-stranded DNA embedded within double-stranded DNA, and that it mostly functions in vitro as a monomer, contrary to what has been described for other cytidine deaminases. This is an important result, because the physiologically- relevant substrate of AID is likely to be transiently formed single-stranded DNA within complex double-stranded structures. We are currently investigating the identity of AID's cofactor and have discovered a novel protein that is under investigation. This protein localizes to replication factories in B cells, and seems to protect B cells from AID activity. We have developed a conditional mouse knockout of this protein to further characterize its function. We are also continuing to examine the role of error-prone DNA polymerases in immunoglobulin hypermutation by the generation of novel mouse strains defective in DNA polymerase Zeta.[unreadable] [unreadable] In addition to these experiments, we have generated lupus-prone, AID-deficient mice to examine the contribution of somatic hypermutation and affinity maturation of antibodies to autoimmune disease. AID deficiency in these mice resulted in a complete alleviation of the associated nephritis, and an increase in survival that, surprisingly, exceeded that of mice that do not secrete any antibodies. These results suggested that absence of pathogenic IgG is not the sole reason for the survival of AID-deficient, lupus-prone mice, but that some protective factor must also contribute. We are now trying to identify the protective factor in these mice and are examining various possibilities such as an increase in regulatory T cells or an increase in protective IgM. Passive transfer experiments into young MRL/lpr mice strongly suggest that it is autoreactive IgM which paradoxically acts to alleviate autoimmunity while its IgG counterpart is pathogenic. In related studies, we also found haplotype insufficiency in AID heterozygotes, which translated into a dramatic decrease in the levels of high-affinity autoantibodies and a concordant decrease in the severity and delay in the onset of lupus nephritis in MRL/lpr mice.